API reference
Base URL https://datda.strategic-innovations.ai. All requests are
JSON. Authenticate with Authorization: Bearer <key>.
POST /v1/scan
Run a scan. Body:
| field | type | notes |
|---|---|---|
| channel | string | input · tool · memory · inter_agent · goal |
| text | string | the untrusted text (for non-tool channels) |
| tool | object | { name, params } (tool channel) |
| context | object | intent, agentId, allowedTools, allowedHosts, confirmed |
Response: verdict, threat_level, score, categories[],
reasons[], detections[], scan_id, engine, usage.
Other endpoints
| method | path | purpose |
|---|---|---|
| GET | /v1/me | identity + entitlement |
| GET | /v1/scans | recent scan history |
| GET/POST/DELETE | /v1/keys | API key management (session) |
| GET/POST | /v1/signatures | custom detection signatures |
| GET | /v1/audit · /v1/audit/verify | tamper-evident audit chain |
| POST | /v1/billing/credits | buy metered scan credits |
| POST | /v1/agent/{id}/quarantine · /release | kill switch (operator only) |
Verdict semantics
threat_level: none → low → medium → high → critical. score is the fused risk in [0,1]. A known-bad signature or a tool-scope breach hard-blocks regardless of score.